# inbox-to-action
[](https://github.com/tarunlnmiit/inbox-to-action/actions/workflows/ci.yml)
[](https://codecov.io/gh/tarunlnmiit/inbox-to-action)
[](https://pypi.org/project/inbox-to-action/)
[](https://www.python.org/)
[](LICENSE)
[](https://github.com/tarunlnmiit/inbox-to-action)
Published on [PyPI](https://pypi.org/project/inbox-to-action/) and listed on the
[Official MCP Registry](https://registry.modelcontextprotocol.io) (`io.github.tarunlnmiit/inbox-to-action`),
[Glama](https://glama.ai/mcp/servers/tarunlnmiit/inbox-to-action) (deployable release, Quality A),
and [Smithery](https://smithery.ai/servers/tarungupta-y12/inbox-to-action) (MCPB bundle).
Registry manifests (`server.json`, `glama.json`) ship in the repo.
[](https://glama.ai/mcp/servers/tarunlnmiit/inbox-to-action)
> One command. Your inbox triaged, summarized, drafted, and turned into tasks โ in a single agentic pass.
## Install
```bash
pip install inbox-to-action # or: pipx install inbox-to-action
uvx inbox-to-action run --mock # zero-install trial (uv)
pip install 'inbox-to-action[mcp]' # + MCP server for Claude Code
docker run --rm ghcr.io/tarunlnmiit/inbox-to-action # MCP server (stdio)
```
Try it with zero setup: `inbox-to-action run --mock` (bundled sample inbox).
๐ **Full documentation โ [docs/](docs/README.md)** โ [install](docs/01-install.md) ยท [providers](docs/02-providers.md) ยท [Gmail OAuth](docs/03-gmail-oauth.md) ยท [multi-account](docs/04-multi-account.md) ยท [integrations](docs/05-integrations.md) ยท [MCP & Skill](docs/06-mcp-and-skill.md) ยท [config](docs/07-config-and-triage.md) ยท [troubleshooting](docs/08-troubleshooting.md) ยท [testing checklist](docs/09-testing-checklist.md). Quick version: [SETUP.md](SETUP.md).
---
## Why this exists
Most people process their inbox with **four** separate tools: an email client to read,
a task manager to capture to-dos, a calendar to block time, and (increasingly) an
AI summarizer to make sense of long threads. Every message gets handled four times.
`inbox-to-action` collapses all four into **one agentic pass**. Run one command and get
a unified triage report, drafted replies saved to Gmail, and extracted tasks โ without
ever leaving the terminal, and **without ever sending an email automatically**.
## ๐ Drafts only โ never sends
This tool **cannot send email**. It requests only the Gmail `readonly` + `compose`
scopes; there is no `send` scope and no send API call anywhere in the codebase
(enforced by a test). Replies are saved as **Gmail drafts** for you to review and send.
Email bodies flow into the LLM prompt, so a hostile email could try to steer its own
classification or a drafted reply (prompt injection). Because every draft is saved for
**human review** and nothing is ever sent automatically, the worst case is a draft you
choose not to send. See [SECURITY.md](SECURITY.md).
## ๐ What leaves your machine
`inbox-to-action` reads your email. **Where your email content goes for classification
depends on the LLM provider you pick** โ and the **default (`openrouter`) is a cloud
provider**, so an out-of-the-box run sends your subjects + bodies to a third party.
| `PROVIDER=` | Email content goes to | Key |
|---|---|---|
| `ollama` | **Nowhere โ fully local** ๐ | none |
| `claude` / `host` | Your existing Claude Code / Anthropic session (keyless) | none |
| `openrouter` **(default)** ยท `openai` ยท `nim` ยท `anthropic` | **Third-party cloud** โ๏ธ | API key |
**Want privacy?** Use `ollama` (local) or `claude` (keyless) so nothing is transmitted
to a third party. `--telegram` / `--todoist` also push subjects/tasks off-box (opt-in).
Full breakdown โ **[PRIVACY.md](PRIVACY.md)**.
> Note: `triage-report.md` and `tasks.md` are written to your working directory and
> contain private email content. If you run inside a git repo, add them to `.gitignore`.
## What it does
1. **Fetches** unread email from Gmail (last 24h by default).
2. **Classifies** each into `action_needed` ยท `fyi` ยท `newsletter` ยท `noise`.
3. **Summarizes** long threads (>500 words) into two lines.
4. **Extracts** tasks with deadlines โ local `tasks.md` (optional Todoist via `--todoist`).
5. **Drafts** replies for `action_needed` mail โ saved as Gmail **drafts**.
6. **Flags** emails that need a calendar block.
Final output: a single **`triage-report.md`** with a section per category, drafted-reply
previews, a tasks summary, and a calendar list.
## Architecture โ the agent loop
The model's own classification of each email drives which tools fire next โ the pipeline
is **not** hardcoded. The same tool functions back the CLI agent and the MCP server.
```mermaid
flowchart TD
CLI[main.py ยท typer] --> AGENT[agent.py ยท agentic loop]
AGENT -->|model picks tools per email| T1[classify_email]
AGENT --> T2[summarize_thread]
AGENT --> T3[extract_tasks]
AGENT --> T4[draft_reply โ Gmail draft]
AGENT --> T5[flag_for_calendar]
T1 & T2 & T3 & T4 & T5 --> LLM[llm_client.py ยท pluggable providers]
LLM --> P1[OpenRouter / Ollama / NIM / OpenAI
OpenAI-compatible HTTP]
LLM --> P2[Anthropic ยท official SDK ยท keyless via ant auth login]
AGENT --> REPORT[report.py โ triage-report.md]
```
```
fetch โ for each email: classify โโฌโ action_needed โ extract_tasks + draft_reply + flag_calendar
โโ fyi / newsletter / noise โ record only
โโ (long thread) โ summarize
โ render triage-report.md
```
## Quick start (2 minutes)
```bash
git clone https://github.com/tarunlnmiit/inbox-to-action.git && cd inbox-to-action
python3 -m venv .venv && source .venv/bin/activate
pip install -e '.[mcp]' # installs the `inbox-to-action` command
cp .env.example .env
```
This installs an `inbox-to-action` console command (and the `python -m
inbox_to_action.mcp_server` entry point used by Claude Code / Glama).
### Free-first: run end-to-end on zero spend
Pick whichever keyless/free path you like โ all run the full pipeline at no cost:
**Option A โ `claude` CLI (keyless, fastest; uses your Claude Code login):**
```bash
PROVIDER=claude inbox-to-action run --mock # no API key; needs `claude` on PATH
```
**Option B โ Ollama (truly keyless, fully local):**
```bash
ollama serve # in another terminal
ollama pull llama3.1
PROVIDER=ollama inbox-to-action run --mock # uses bundled sample inbox
```
**Option C โ OpenRouter free model (free signup key):**
```bash
# put OPENROUTER_API_KEY in .env (free models, $0 spend)
inbox-to-action run --mock # default PROVIDER=openrouter
```
**Option D โ inside Claude Code (keyless, Claude Code is the LLM):** see below.
`--mock` uses the bundled sample inbox so you can see a full report with **zero Gmail
setup**. Drop `--mock` once you've authorized Gmail. Free OpenRouter models are often
rate-limited; the client auto-rotates a fallback list and retries with backoff.
### Real inbox
```bash
# 1. Create OAuth credentials in Google Cloud Console (Desktop app),
# download client_secret.json into the project, then:
inbox-to-action auth # one-time consent (read + compose only)
inbox-to-action run --since 24h --no-drafts # safe first pass: report only, no writes
inbox-to-action run --since 24h # triage the last day (creates Gmail drafts)
inbox-to-action run --since 3d --max 40 --todoist
```
- `--no-drafts` โ classify, summarize, extract tasks, write the report, but create
**no** Gmail drafts. Recommended for a first run.
- `--max N` โ cap emails per account (default 25) to bound cost/volume.
- Automated **no-reply** senders (security alerts, notifications) never get a drafted
reply โ the report notes them instead.
### Telegram summary (`--telegram`)
Push a concise summary to your phone after each run โ counts, action-needed subjects
(with draft-ready status), extracted tasks, and a link to your Gmail Drafts.
```bash
# 1. In Telegram, message @BotFather โ /newbot โ copy the bot token.
# 2. Message your new bot once (say "hi"), then open:
# https://api.telegram.org/bot/getUpdates โ copy "chat":{"id": ...}.
# 3. Put both in .env:
# TELEGRAM_BOT_TOKEN=... TELEGRAM_CHAT_ID=...
inbox-to-action run --since 24h --telegram
```
Off by default (opt-in flag). A send failure never breaks the run.
**Privacy:** this sends email subjects + extracted tasks to Telegram's servers (into
your own chat). It's notification only โ it never sends email.
### Multiple accounts (Gmail + Google Workspace)
Declare accounts in `config.json` โ one merged report, each email tagged with its
account. Personal Gmail and Workspace both use the Gmail path (Workspace may need
your admin to allow the OAuth app).
```json
{
"accounts": [
{ "id": "personal", "kind": "gmail", "label": "Personal Gmail" },
{ "id": "work", "kind": "gmail", "label": "Workspace" }
]
}
```
```bash
inbox-to-action auth --account personal # authorize each account once
inbox-to-action auth --account work
inbox-to-action run --since 24h # fetches + triages across all accounts
```
Multiple personal Gmail accounts can reuse one `client_secret.json` โ each gets its
own cached token (`~/.config/inbox-to-action/tokens/.json`). With no `accounts`
block, the tool uses a single default Gmail account (backwards compatible).
## Use inside Claude Code (keyless)
When run inside Claude Code, **Claude Code is the LLM** โ no provider key needed.
Two integration paths ship in this repo:
### MCP server
Exposes IO-only tools (`fetch_emails`, `save_gmail_draft`, `append_tasks`, `write_report`).
Claude Code does the classify/summarize/extract/draft reasoning itself and calls these.
```bash
# after `pip install -e '.[mcp]'`
claude mcp add inbox-to-action -- python -m inbox_to_action.mcp_server
```
This is the same stdio server that MCP registries (e.g. [Glama](https://glama.ai)) build from
the bundled `Dockerfile` (`CMD python -m inbox_to_action.mcp_server`).
### Skill
Copy `skills/inbox-to-action/` into your Claude Code skills directory, then type
`/inbox-to-action`. The skill instructs Claude Code to fetch, reason, draft, and write
the report โ keyless.
## Anthropic (keyless via `ant auth login`)
The Anthropic provider uses the official SDK with a zero-arg client, so it picks up your
`ant auth login` OAuth profile โ **no `ANTHROPIC_API_KEY` required**:
```bash
ant auth login
PROVIDER=anthropic inbox-to-action run --mock # default model: claude-opus-4-8
```
## Configuration
All keys live in `.env` (`.env.example` is committed). Switch providers with `PROVIDER`:
`openrouter` (default) ยท `ollama` ยท `nim` ยท `openai` ยท `anthropic` ยท `claude` ยท `host`.
### Configure triage (make it *yours*)
The default buckets are generic โ newsletters and job alerts are treated as no-action.
Override that with `config.json` (copy `config.example.json`). Two layers:
- **`rules`** โ deterministic `field โ category` overrides applied **before** the LLM
(fast, free, exact). First match wins. `field` โ `sender | subject | body | any`.
- **`triage_instructions`** โ freeform guidance injected into the classifier prompt for
nuance the model interprets.
```json
{
"triage_instructions": "I'm job hunting in ML/AI โ treat relevant job alerts as action_needed.",
"rules": [
{ "field": "sender", "contains": "hirist.tech", "category": "action_needed" },
{ "field": "subject", "contains": "invoice", "category": "noise" }
]
}
```
```bash
cp config.example.json config.json # edit to taste (config.json is gitignored)
inbox-to-action run --since 24h # auto-loads ./config.json
inbox-to-action run --config /path/to/other.json
```
Quick override without a file: `TRIAGE_INSTRUCTIONS="treat job alerts as action_needed"`.
## Tests
```bash
pytest --cov=. # 100+ tests, ~89% coverage, incl. the never-send security test
```
## Docs
- **[docs/](docs/README.md)** โ full guides with screenshots: install, every LLM provider, Gmail OAuth, multi-account, integrations, MCP & Skill, config, troubleshooting, testing checklist.
- [SETUP.md](SETUP.md) โ 5-minute quickstart.
- [PRIVACY.md](PRIVACY.md) โ what leaves your machine, per provider.
- [SECURITY.md](SECURITY.md) โ never-send invariant + vulnerability disclosure.
- [CHANGELOG.md](CHANGELOG.md) โ version history.
- [CONTRIBUTING.md](CONTRIBUTING.md) โ dev setup + the never-send rule.
- [CLAUDE.md](CLAUDE.md) โ project map for Claude Code.
## Built with
This project demonstrates the contract skills:
- **Agentic orchestration** โ model-driven, per-email tool selection (no hardcoded pipeline).
- **Function calling** โ typed tool schemas (`agent.TOOL_SCHEMAS`) shared by the CLI agent and MCP server.
- **Multi-API integration** โ Gmail + LLM + Todoist in one flow.
- **Pluggable LLM providers** โ one `llm_client` swaps OpenRouter / Ollama / NIM / OpenAI / Anthropic.
- **Claude Code integration** โ first-class MCP server **and** Skill, both keyless.
## License
MIT โ see [LICENSE](LICENSE).